GET

/v1/schemas/fal/audio/fal-ai/minimax/speech-2.8-hd

Accept:

kind inputactivity audiohash 47394979…extracted Oct 17, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `35bb95a65e534ac28f178bc39624abd4083c4610248e05161ab5c1c5c62df667` { "fetchedAt": 1792213207, "extractorVersion": "1", "verifiedAt": null }

Request body

9 properties · 1 required · 5 $defs

property	type	description	constraints
prompt*	string	Text to convert to speech. Use `<#x#>` for pauses (x = 0.01-99.99 seconds). Supports interjection tags: `(laughs)`, `(sighs)`, `(coughs)`, `(clears throat)`, `(gasps)`, `(sniffs)`, `(groans)`, `(yawns)`.	length ≥ 1 · length ≤ 10000
voice_setting	VoiceSetting	Voice configuration settings
language_boost	enum	Enhance recognition of specified languages and dialects	Chinese \| Chinese,Yue \| English \| Arabic \| Russian \| Spanish · +32 more · nullable
voice_modify	VoiceModify	Voice modification settings for Speech 2.8 models.	nullable
output_format	enum	Format of the output content (non-streaming only)	url \| hex · default: "hex"
normalization_setting	LoudnessNormalizationSetting	Loudness normalization settings for the audio
enable_safety_checker	boolean	Authorized callers may set false to request relaxed screening.	default: true
audio_setting	AudioSetting	Audio configuration settings
pronunciation_dict	PronunciationDict	Custom pronunciation dictionary for text replacement	nullable

Types

5 $defs

AudioSettingobject

bitrate	enum	Bitrate of generated audio	32000 \| 64000 \| 128000 \| 256000 · default: 128000
format	enum	Audio format	mp3 \| pcm \| flac · default: "mp3"
channel	enum	Number of audio channels (1=mono, 2=stereo)	1 \| 2 · default: 1
sample_rate	enum	Sample rate of generated audio	8000 \| 16000 \| 22050 \| 24000 \| 32000 \| 44100 · default: 32000

LoudnessNormalizationSettingobject

target_range	number	Target loudness range in LU (default 8.0)	0 ≤ n ≤ 20 · default: 8
target_peak	number	Target peak level in dBTP (default -0.5).	-3 ≤ n ≤ 0 · default: -0.5
enabled	boolean	Enable loudness normalization for the audio	default: true
target_loudness	number	Target loudness in LUFS (default -18.0)	-70 ≤ n ≤ -10 · default: -18

PronunciationDictobject

tone_list

string[]

List of pronunciation replacements in format ['text/(pronunciation)', ...]. For Chinese, tones are 1-5. Example: ['燕少飞/(yan4)(shao3)(fei1)']

VoiceModifyobject

Voice modification settings for Speech 2.8 models.

pitch	integer	Pitch adjustment in semitones. Range: -100 to 100. Positive values raise pitch, negative values lower it.	-100 ≤ n ≤ 100 · default: 0
timbre	integer	Timbre adjustment. Range: -100 to 100. Affects the tonal quality of the voice.	-100 ≤ n ≤ 100 · default: 0
intensity	integer	Intensity/energy of the voice. Range: -100 to 100. Higher values create more energetic speech.	-100 ≤ n ≤ 100 · default: 0

VoiceSettingobject

pitch	integer	Voice pitch (-12 to 12)	-12 ≤ n ≤ 12 · default: 0
vol	number	Volume (0-10)	0.01 ≤ n ≤ 10 · default: 1
english_normalization	boolean	Enables English text normalization to improve number reading performance, with a slight increase in latency	default: false
speed	number	Speech speed (0.5-2.0)	0.5 ≤ n ≤ 2 · default: 1
emotion	enum	Emotion of the generated speech	happy \| sad \| angry \| fearful \| disgusted \| surprised · +1 more · nullable
voice_id	string	Predefined voice ID to use for synthesis	length ≥ 1 · default: "Wise_Woman"

Validate a payload

before you spend tokens

shell

$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"fal","endpointId":"fal-ai/minimax/speech-2.8-hd","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}